What is the difference between CLAUDE.md and .llmdocs?

CLAUDE.md is a concise, high-level overview (the 'Map') that is always loaded for context, while .llmdocs contains detailed, per-concept files (the 'Territory') loaded only on demand.

Does this skill overwrite my existing documentation?

No, it is designed to update documentation incrementally. It analyzes existing content and asks for user validation before creating, modifying, or removing any documentation.

Why does the skill avoid using prose paragraphs?

Large Language Models process structured data—such as headers, tables, and code blocks—more efficiently than long-form prose when seeking specific technical information.

How does the skill know what needs to be updated?

The skill reviews your recent conversation history, inspects git diffs, and explores the directory structure to identify which components or logic have changed during the session.

LLM Documentation Sync

Name: LLM Documentation Sync
Author: jimweller

byjimweller

0•

学习与文档

Synchronizes project documentation and AI context files with recent codebase changes to ensure accurate LLM performance.

This skill maintains a dual-layer documentation system designed specifically for AI-assisted development. It manages a concise CLAUDE.md file that serves as a high-level project map and a modular .llmdocs/ directory for deep technical dives into architecture, APIs, and data models. By analyzing session history and git diffs, the skill identifies precisely what needs updating, enforcing strict formatting rules that optimize readability for Large Language Models while preventing context window bloat.

主要功能

01Generates modular, concept-specific documentation within a dedicated .llmdocs directory.

02Maintains a high-level CLAUDE.md project map under 150 lines for instant context.

03Analyzes git diffs and conversation history to automatically detect necessary updates.

04Enforces LLM-optimized formatting using structured lists and tables instead of verbose prose.

05Provides a summary of all modified sections and files for easy user verification.

060 GitHub stars

使用场景

01Ensuring Claude has up-to-date context on project conventions and command structures.

02Standardizing documentation across a team to improve AI coding efficiency and accuracy.

03Updating project guides and architecture maps immediately after a major feature refactor.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jimweller/davit-dataops-dotfiles update-llm-docs

For use in Claude.ai and ChatGPT

主要功能

01Generates modular, concept-specific documentation within a dedicated .llmdocs directory.

02Maintains a high-level CLAUDE.md project map under 150 lines for instant context.

03Analyzes git diffs and conversation history to automatically detect necessary updates.

04Enforces LLM-optimized formatting using structured lists and tables instead of verbose prose.

05Provides a summary of all modified sections and files for easy user verification.

060 GitHub stars

使用场景

01Ensuring Claude has up-to-date context on project conventions and command structures.

02Standardizing documentation across a team to improve AI coding efficiency and accuracy.

03Updating project guides and architecture maps immediately after a major feature refactor.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jimweller/davit-dataops-dotfiles update-llm-docs

For use in Claude.ai and ChatGPT